Only one-quarter of AI use cases in customer service produce a positive return on investment, according to a Gartner analysis of 432 use cases released last month. Another quarter deliver negative returns, and 42% have unclear ROI, with support leaders saying they simply don't know the value produced. Just 11% of customer support use cases break even.
Despite those results, more than three-quarters of customer service leaders plan to increase AI investment in 2026. The disconnect matters because customer service leads AI adoption across the enterprise - teams are pursuing an average of nearly five AI use cases and committing about 13% of their functional budget to the technology.
Pressure from the top
The gap between investment and measurable value stems from a top-down approach that doesn't start with a customer problem, experts said. Executives are directing support teams to deploy AI without defining what success looks like.
"Too many AI rollouts begin with pressure to demonstrate a credible AI strategy to the board, rather than with a clearly defined business problem," said Julie Geller, principal research director at Info-Tech Research Group.
That pressure is becoming personal for support leaders. More than half - 56% - expect to have their incentives tied directly to AI outcomes in 2026, even though those outcomes remain hard to prove.
Containment is not resolution
Many businesses expect cost savings through workforce reduction, with AI agents handling the easy questions that customer service representatives once answered. But Gartner found the rate of organizations increasing head count equals the rate reducing it - one-quarter report workforce growth and one-quarter report reductions.
As companies adopt AI, they also need new specialized roles to manage the technology. And the common goal of containment - keeping customers away from human agents - misses the point if it doesn't actually help the customer.
"Containment is also too often mistaken for success," Geller said. "Delaying contact with a human agent is not the same as resolving the customer's problem. The real test is much simpler: did the customer get what they needed, with less effort?"
Gartner's findings align with a recent report from Forethought AI Agents by Zendesk. While 70% of organizations have rolled out AI in customer experience, only a small percentage are producing value in the form of improved outcomes and ROI.
Why this matters for customer support professionals
If your team is pursuing multiple AI use cases, you need to define the business problem before the rollout - not after. Ask what customer pain point the technology solves, and measure whether customers actually get what they need with less effort. Track ROI per use case rather than treating AI investment as a single initiative, and be prepared to show results if your compensation becomes tied to AI outcomes in the coming year.
Your membership also unlocks:
